Interaction of 14-3-3 Protein with Regulator of G Protein Signaling 7 Is Dynamically Regulated by Tumor Necrosis Factor-α*

Regulators of G protein signaling (RGS) constitute a family of proteins with a conserved RGS domain of ∼120 amino acids that accelerate the intrinsic GTP hydrolysis of activated Gαi and Gαq subunits. The phosphorylation-dependent interaction of 14-3-3 proteins with a subset of RGS proteins inhibits their GTPase-accelerating activity in vitro. The inhibitory interaction between 14-3-3 and RGS7 requires phosphorylation of serine 434 of RGS7. We now show that phosphorylation of serine 434 is dynamically regulated by TNF-α. Cellular stimulation by TNF-α transiently decreased the phosphorylation of serine 434 of RGS7, abrogating the inhibitory interaction with 14-3-3. We examined the effect of 14-3-3 on RGS-mediated deactivation kinetics of G protein-coupled inwardly rectifying K+ channels (GIRKs) in Xenopusoocytes. 14-3-3 inhibited the function of wild-type RGS7, but not that of either RSG7P436R or RGS4, two proteins that do not bind 14-3-3. Our findings are the first evidence that extracellular signals can modulate the activity of RGS proteins by regulating their interaction with 14-3-3.

G protein-coupled receptors activate heterotrimeric G proteins by enhancing GDP release and binding of GTP to the G␣ subunit. By accelerating the inactivation of GTP-bound G␣ i and G␣ q subunits, the "regulator of G protein signaling" (RGS) 1 proteins negatively regulate the strength and duration of G protein signaling. RGS proteins act catalytically; a single molecule binds and inactivates multiple GTP-bound G␣ subunits. Functional specificity and temporal activity of RGS proteins arises from patterns of expression, transcriptional regulation, subcellular localization, and posttranslational modification (1). Whereas some RGS proteins are ubiquitously expressed, others are temporally and spatially restricted to certain tissues and developmental stages or are induced by cellular activation (reviewed in Refs. 2 and 3). Interaction with other cytoplasmic proteins affects the stability, function, and subcellular localiza-tion of RGS proteins (reviewed in Refs. 4 and 5).
We have previously shown that binding to 14-3-3 modulates the activity of certain RGS proteins, including RGS7, in vitro (6). These RGS proteins contain a conserved SYP motif with upstream basic residues located in one of the three contact sites formed between G␣ i and the RGS domain. Binding of 14-3-3 to the SYP motif of RGS7 requires the phosphorylation of the serine residue at position 434. In this study, we demonstrate that TNF-␣ inhibits the phosphorylation of serine 434 and the interaction of RGS7 with 14-3-3. RGS7 is highly expressed in the mouse brain, with a substantial fraction associated with 14-3-3. Treatment of mice with TNF-␣ completely abrogated the interaction of RGS7 with 14-3-3. We examined the effect of 14-3-3 on RGS-mediated deactivation kinetics of G proteincoupled inwardly rectifying K ϩ channels (GIRKs) in a Xenopus co-expression system (7,8). Microinjected 14-3-3 protein slowed the deactivation of GIRKs in the presence of wild-type RGS7, but had no effect in the presence of either RSG7 P436R or RGS4, two proteins that do not associate with 14-3-3. Our findings demonstrate that the activity of RGS7 is regulated by extracellular signals.

EXPERIMENTAL PROCEDURES
Plasmids-FLAG-tagged versions of human RGS7, RGS3, and 14-3-3 have been described previously (6). FLAG-tagged RGS4 and RGS8 were cloned from human cDNA libraries generated by standard cloning procedures. Site-directed mutagenesis was used to insert mutations in RGS7. Point mutations were verified by sequence analysis. M1 and M2 muscarinic receptors and the GIRK constructs were kind gifts of Dr. Silvio Gutkind, Dr. Melanie Mark, and Dr. Ernest Peralta.
Generation of a Phospho-specific Phosphoserine 434 Antiserum-A phosphopeptide corresponding to residues 429 -451 of human RGS7 (acetyl-CLMKSDpSYPRFIRS-amide) was conjugated to keyhole limpet hemocyanin and used to generate polyclonal antiserum (Quality Controlled Biochemicals, Hopkinton, MA). The rabbit antiserum was preabsorbed three times with the corresponding non-phosphopeptide coupled to an agarose column. Specificity and titer were monitored by Western blotting and enzyme-linked immunosorbent assay, respectively, using phospho-and non-phosphopeptide conjugated to BSA. Phospho-specific antiserum was affinity-purified by passing the flowthrough of the third preabsorption over a phosphopeptide column.
Immunoprecipitation and Immunoblotting-Immunoprecipitation experiments were performed as described (6). Briefly, HEK 293T cells were transiently transfected by the calcium phosphate method. After incubation for 24 h, untreated cells and cells treated with 10 ng/ml TNF-␣ were washed twice and lysed in 1% Triton X-100 lysis buffer. After centrifugation (15,000 ϫ g, 15 min, 4°C), cell lysates containing equal amounts of total protein were incubated for 1 h at 4°C with the appropriate antibody followed by incubation with 40 l of protein G-Sepharose (for mouse monoclonal antibodies) or protein A-Sepharose (for rabbit polyclonal antisera) (Amersham Biosciences) for ϳ3 h. Following extensive washing with lysis buffer, proteins samples were separated by SDS-PAGE, transferred to nitrocellulose membrane, and following primary and secondary antibody incubation, detected by chemiluminescence.
Pull-down Assay-HEK 293T cells were transiently transfected with FIG. 1. Generation of a phospho-specific RGS7 antiserum. A, structural elements of RGS7 include a DEP domain (dishevelled/EGL-10/ pleckstrin), a G␥-like (GGL) domain, an RGS (regulator of G protein signaling) domain, and a 14-3-3 binding motif consisting of (K/E)-(K/R)-D-pS-Y-P (6). B, the RGS7 phosphopeptide CLMKSDpSYPRFIRS was used to generate an antiserum that specifically recognizes RGS7 when phosphorylated on serine 434. Western blotting analysis using the phospho-specific RGS7 antiserum against dilutions of phosporylated RGS7 peptide coupled to BSA, non-phosphorylated RGS7 peptide coupled to BSA, and BSA alone. C, PKC ␣ catalyzes the incorporation of phosphate into a truncation of RGS7 fused to maltose-binding protein (MBP.RGS7 315-469 ). D, GST, lacking potential PKC phosphorylation sites, is not phosphorylated by PKC ␣ in vitro. Since the RGS7 315-469 truncation contains several potential PKC sites (Ser 434 , Ser 442 , Ser 456 , Thr 458 , and Ser 468 ), both wild-type GST.RGS7 315-469 and the S434D mutant of GST.RGS7 315-469 are phosphorylated by PKC ␣ with no discernible difference by autoradiography. The faint band at 83 kDa represents autophosphorylated PKC ␣. Equal loading of recombinant proteins was verified by staining the protein gel with Coomassie Blue, as depicted in the lower panel. E, the phospho-specific RGS7 antiserum recognizes wild-type MBP.RGS7 315-469 , but not the S434D mutation of MBP.RGS7 315-469 after phosphorylation with PKC ␣. Equal loading of recombinant proteins was verified by staining the protein gel with Coomassie Blue, as depicted in the lower panel. F, the phospho-specific RGS7 antiserum immunoprecipitates phosphorylated, but not non-phosphorylated, GST.RGS7 315-469 nor GST alone. Precipitated proteins were detected with anti-GST antisera (Amersham Biosciences).

FIG. 2. TNF-␣ regulates phosphorylation of serine 434 of RGS7 and binding to 14-3-3.
A, TNF-␣ reduces serine 434 phosphorylation of RGS7. HEK 293T cells were transfected with FLAG-tagged RGS7 or RGS3. Despite an increase of F.RGS7 protein levels after a 1-h incubation with TNF-␣ (left panel), virtually no RGS7 was precipitated by the phospho-specific RGS7 antiserum (right panel), demonstrating that TNF-␣ reduces phosphorylation of RGS7 on serine 434. Precipitated proteins were detected with the FLAG-specific M2 antibody. The phospho-specific RGS7 antiserum is highly specific for RGS7 and did not precipitate FLAG-tagged RGS3, a control protein that contains the SYP motif and binds 14-3-3 (right panel). B, TNF-␣ abrogates binding of 14-3-3 to RGS7. HEK 293T cells were transfected with FLAG-tagged RGS7 and Myc-tagged 14-3-3. Despite progressively increasing F.RGS7 proteins levels (lysates, top panel), the amount of F.RGS7 precipitated by the phospho-specific RGS7 antiserum drastically declines after 30 min of TNF-␣ stimulation (IP, middle panel); precipitated proteins were detected with the FLAG-specific M2 antibody. The diminished serine 434 phosphorylation observed at 30 min coincides with decreased co-precipitation of 14-3-3 with RGS7 (IP, bottom panel). Following precipitation of F.RGS7 by FLAG-specific M2 antibody, co-precipitating Myc.14-3-3 was detected using an anti-Myc antibody. C, binding of RGS7 to recombinant GST.14-3-3 is diminished by TNF-␣. HEK 293T cells were transfected with a vector control or FLAG-tagged RGS7. Cell lysates were incubated with GST.14-3-3 immobilized on glutathione-Sepharose. Bound FLAG-tagged RGS7 was detected by Western blot analysis, using the FLAG-specific M2 antibody. GST.14-3-3 immobilized RGS7 from resting cells, but failed to bind significant amounts of RGS7 from cells stimulated with TNF-␣. D, TNF-␣ transiently decreases association of 14-3-3 to RGS7. HEK 293T cells, transfected with FLAG-tagged RGS7 and Myc-tagged 14-3-3, were stimulated with TNF-␣ for the indicated times. Following precipitation of F.RGS7 by FLAG-specific M2 antibody, co-precipitating Myc.14-3-3 was detected using an anti-Myc antibody. Decreased binding of 14-3-3 to RGS7 is evident at 30 and 60 min. E, TNF-␣ decreases serine 434 phosphorylation of RGS7. HEK 293T cells were transfected with RGS7 tagged with YFP. Cellular lysates were assessed by Western blot analysis with phosphoserine-specific antiserum to detect dephosphorylation at serine 434 induced by TNF-␣. A longer form of RGS7 was used to avoid interference from a nonspecific 60-kDa band arising from immunoblotting with the phosphoserine-specific antiserum.
loading of GST.14-3-3 was confirmed by Coomassie Blue staining of the gels.
Immunoprecipitation of Mouse Brain-For preparation of brain protein extracts, female BALB/c mice (20 g in body weight, Charles River) were injected intravenously in the tail vein with solvent or 20 g of murine TNF-␣. After 4 h, the mice were sacrificed and brains were removed and homogenized essentially as described (1). Following centrifugation and ultracentrifugation (100,000 ϫ g, 4°C, 30 min), the supernatant was subjected to several preclearing steps, divided into two fractions, and immunoprecipitated with specific anti-RGS7 antiserum (rabbit polyclonal antiserum generated against MBP-RGS7 170 -469 and affinity-purified with GST-RGS7 315-469 ) and control antibody, followed by incubation with protein G-Sepharose. Resulting precipitates were subjected to immunoblot analysis with anti-14-3-3 monoclonal antibody (Santa Cruz) followed by incubation with horseradish peroxidasecoupled secondary antiserum and enhanced chemiluminescence.
Preparation and Injection of Oocytes-Xenopus laevis frogs were anesthesized (4 mM aminobenzoic acid ethyl ester for 25 min), and oocytes were isolated by partial ovarectomy. Subsequently, oocytes were manually dissected from ovarian lobes and defolliculated by treatment with 1 mg/ml collagenase type A (Roche Molecular Biochemicals, Mannheim, Germany) in a Ca 2ϩ -free hypotonic solution (82.5 mM NaCl, 2 mM KCl, 1 mM MgCl 2 , 5 mM HEPES, pH 7.4). Oocytes were washed, and transferred to ND96 solution (96 mM NaCl, 2 mM KCl, 1.8 mM CaCl 2 , 1 mM MgCl 2 , 5 mM HEPES, 2.5 mM sodium pyruvate, pH 7.4, supplemented with 50 ng/ml gentamicin and 0.5 mM theophylline. cRNAs were synthesized in vitro using the mMessage Machine kit (Ambion). 12-24 h after isolation, stage V-VI oocytes were injected with 30 nl of water containing 1 ng of cRNA for the M2 receptor, 0.1 ng for GIRK1 and GIRK2, and 10 ng for the different RGS proteins. Control oocytes were injected with water instead of RGS-cRNAs. Voltage clamp experiments were performed 2-4 days after injection of cRNAs. Recombinant proteins were injected into the oocytes 2-12 h prior to the experiments. 30 nl of GST (0.5 mg/ml) and GST.14-3-3 (1 mg/ml) were injected resulting in a final cytosolic concentration of ϳ500 nM.
Electrophysiology-Standard two-electrode voltage clamp was used to assess the RGS protein function in vivo. Whole cell currents of oocytes were recorded using the Turbo TEC 03X voltage/current clamp amplifier (NPI Electronic, Tamm, Germany). Microelectrodes were pulled on a vertical puller (Physiologisches Institut, Freiburg, Germany) from borosilicate glass capillaries (Clark Instruments, Reading, United Kingdom) and had resistances of 0.5-2 M⍀ when filled with 2 M KCl solution. Experiments were performed at 20°C. All whole cell voltage clamp experiments were conducted in ND96 solution or a modified ND96-K80 solution in which 80 mM of NaCl were replaced with KCl. Acetycholine (1 M) and atropine (1 M) were added to the ND96-K80 solution to activate and antagonize the M2 receptor. Oocytes were clamped at 0 mV, and voltage ramps from Ϫ100 to ϩ50 mV were applied every 1.2 s. Time constants for current decrease after removal of acetycholine were fitted with a single exponential (Origin, Additive, Germany). Data are presented as original recordings and as mean values Ϯ S.E. (n ϭ number of experiments). Unpaired Student's t test was used for statistical analysis. A p value of Ͻ0.05 was accepted to indicate statistical significance.
serum. To affinity-purify antiserum that specifically recognized phosphorylated serine 434, phosphorylated and non-phosphorylated peptides were covalently coupled to agarose. After several rounds of positive and negative selection, the phosphospecific serine 434 antiserum recognized less than 1 pmol of RGS7-derived peptide, but did not bind non-phosphorylated peptide nor BSA (Fig. 1B). Serine 434 can be phosphorylated by several protein kinase C isozymes (PKC ␣, ␤, ␥, ⑀, , ; scansite.mit.edu/, Motif Scanner at low stringency). To test whether phosphorylated RGS7 was specifically recognized by the phosphoserine 434-specific antiserum, the RGS7 truncation containing the RGS domain (RGS7 315-469 ) was fused to either maltose-binding protein (MBP) or GST and phosphorylated with PKC ␣ (Fig. 1, C-E). Due to the presence of multiple potential PKC phosphorylation sites (Ser 442 , Ser 456 , Thr 458 , Ser 468 ), the overall level of phosphorylation is unchanged by the mutation of serine 434 to aspartic acid (Fig. 1D). However, the phospho-specific antiserum specifically recognized phosphorylated, but not non-phosphorylated RGS7 315-469 nor the S434D mutation of RGS7 315-469 (Fig. 1E). Furthermore, the antiserum specifically immunoprecipitates phophorylated RGS7 315-469 fused to GST, but not a non-phosphorylated GST-RGS7 315-469 fusion protein nor GST alone (Fig. 1F). These experiments demonstrate the specificity of the antiserum for phosphorylated serine 434 of recombinant RGS7 in both immunoblotting and immunoprecipitation.
TNF-␣ Causes Dephosphorylation of Serine 434 of RGS7 and Dissociation with 14-3-3-We have previously shown that binding to 14-3-3 modulates the activity of certain RGS proteins, including RGS7 (6). To determine whether TNF-␣ mediates the accumulation of functionally active RGS7, we monitored serine 434 phosphorylation of RGS7 in transfected HEK 293T cells.
Although lysates of TNF-treated cells contain more RGS7 than untreated cells ( Fig. 2A, left panel), precipitation of RGS7 by the phospho-specific antiserum is substantially decreased ( Fig.  2A, right panel). The phospho-specific antiserum was used to precipitate phosphoserine 434-containing RGS7; precipitated protein was then detected using the FLAG-specific M2 antibody. After 30-min stimulation with TNF-␣, the accumulation of RGS7 is accompanied by decreased phosphorylation of serine 434 and marked dissociation with 14-3-3 (Fig. 2B). These results are supported by pull-down experiments using a recombinant GST.14-3-3 fusion protein immobilized on glutathione-Sepharose. Treatment of cells with TNF-␣ both rapidly and greatly reduced the amount of RGS7 retained on a GST.14-3-3 column, compared with untreated cells (Fig. 2C). Fig. 2D demonstrates the transient capacity of TNF-␣ to dissociate RGS7 and 14-3-3. In transfected HEK 293T cells, binding of RGS7 to 14-3-3 is inhibited after 30-and 60-min treatment with TNF-␣, but reassociation is evident at 120 min. The dephosphorylation of RGS7 induced by TNF-␣ occurs rapidly. As shown in Fig. 2E, treatment with TNF-␣ reduced serine 434 phosphorylation of RGS7 at 15 and 30 min. These results indicate that TNF-␣ induces dephosphorylation of serine 434 and transiently re-

FIG. 5. 14-3-3 inhibits the RGS7-mediated acceleration of GIRK deactivation kinetics in Xenopus oocytes.
Xenopus oocytes were microinjected with cRNA encoding for the G protein-activated inwardly rectifying K ϩ channels subunits GIRK1 and GIRK2, the M2 acetylcholine receptor (M2 AChR), and RGS7 (or water in control oocytes) as indicated. GIRK currents were recorded using the two-electrode voltage-clamp technique. A, current-voltage relations of an oocyte expressing GIRK1/2 and the M2 AChR in the presence of different extracellular solutions: control, ND96; 80 mM K ϩ , ND96-K80; and ACh 1 M, acetylcholine (1 M) added to ND96-K80. B, representative experiment of an oocyte expressing GIRK1/2 and M2 AChR. Oocytes were clamped at 0 mV, and voltage ramps from Ϫ100 to ϩ50 mV were applied every 1.2 s. C, deactivation kinetics of control oocytes and oocytes expressing RGS7 (ϩRGS7). Deactivation time constants were derived from single exponential fits of the GIRK current deactivation phase from control oocytes or oocytes expressing RGS7 following acetylcholine removal (holding potential Ϫ90 mV). D, comparison of deactivation time constants of control oocytes and oocytes expressing RGS7. Expression of RGS7 significantly decreased the deactivation time constant of acetylcholine-evoked currents. E, microinjection of GST.14-3-3 protein, but not GST, inhibited the RGS7-mediated acceleration of GIRK1/2 deactivation after withdrawal of acetylcholine. leases RGS7 from its inhibitory interaction with 14-3-3.
TNF-␣ Blocks 14-3-3 Binding to RGS7 in Vivo-We previously reported that TNF-␣ inhibits the degradation of RGS7 (1), resulting in the accumulation of RGS7 in the brains of mice treated with TNF-␣ or endotoxin, a stimulator of TNF-␣ release and formation. In the brain, a significant fraction of RGS7 appears to be phosphorylated on serine 434 and complexed with 14-3-3, conditions known to inhibit the in vitro GAP activity of RGS7 (6). To examine the dual effects of TNF-␣ in vivo, we examined the interaction of RGS7 with 14-3-3 in mice treated with TNF-␣. As shown in Fig. 4A, co-immunoprecipitation of 14-3-3 with RGS7 was evident in brains of solventtreated animals. This interaction was completely abrogated in mice treated with TNF-␣ (Fig. 4A, lower panel), although levels of 14-3-3 expression remained unchanged (Fig. 4A, upper panel). Binding of the adaptor protein 14-3-3 strictly depends upon a phosphorylated 14-3-3 binding site. As demonstrated in Fig.   4B, the effect of TNF-␣ can be mimicked by calf intestine alkaline phosphatase. Treatment of immunoprecipitated RGS7 with calf intestine alkaline phosphatase markedly decreased 14-3-3 binding, indicating that the interaction between RGS7 and 14-3-3 can be abrogated by phosphatases. It is likely that TNF-␣ triggers the activation of a currently unknown phosphatase that is absent or inactive in resting cells.
14-3-3 Modulates the Deactivation Kinetics of GIRKs Mediated by RGS7-GIRKs open upon binding to G␤␥ subunits released by the activation of pertussis toxin-sensitive G␣ i /G␣ ocoupled receptors (9). Activation and deactivation kinetics of GIRK channels control the onset and termination of postsynaptic hyperpolarizing currents in neurons and cardiac atrial cells. The kinetics of neuronal and atrial GIRK channel activation and deactivation are markedly faster than those observed in heterologously expressed channels. Several RGS proteins, including RGS1, RGS3, RGS4, RGS7, RGS8, and RGS9, increase the rate at which GIRK channels close following agonist removal, presumably by accelerating the GTPase activity of G␣ i and thereby sequestering G␤␥ (7, 8, 10 -14). In the present study, we examined the effect of 14-3-3 on RGS7-mediated changes in deactivation kinetics of heterologously expressed GIRK1/2 channels. Oocytes were injected with water or RGS7 mRNA in combination with the muscarinic M2 acetylcholine receptor (M2 AChR) and GIRK1/GIRK2 to mimic the heteromultimeric state of native neuronal GIRK channels. Since the deactivation kinetics of GIRK channels are not affected by the ratio of G protein-coupled receptors to ion channel, this experimental setting permits the quantification of RGS protein function (15). As reported previously (7), ACh-evoked GIRK currents recorded from Xenopus oocytes in the absence of RGS proteins are deactivated with a time course best described by a single time constant of ϳ20 s (Fig. 5). In the presence of RGS7, the deactivation kinetics after removal of acetylcholine was significantly accelerated with a time constant of 13 s (Fig. 5, C  and D). Microinjection of bacterially expressed and affinity purified recombinant GST.14-3-3 protein, but not of GST alone, significantly reduced the RGS7-mediated deactivation of GIRK currents (Fig. 5E). However, 14-3-3 did not influence GIRK channels kinetics in the absence of RGS proteins (data not shown). These findings indicate that interaction with 14-3-3 inhibits the activity of RGS7 in vivo.
A Conserved SYP/S/T Motif within the RGS Domain Interacts with 14-3-3-To confirm the specificity of the 14-3-3-induced inhibition of RGS7 activity in oocytes, we examined the effect of 14-3-3 on RGS4, a known accelerator of GIRK current deactivation that does not interact with 14-3-3. RGS4 accelerated the deactivation of GIRK1/2 in microinjected oocytes (Fig.  6, A and B). This RGS4-induced acceleration of GIRK channel deactivation was unaffected by co-injection with GST or GST.14-3-3 (Fig. 6C). The RGS domain of RGS4 contains an SYR motif that does not bind 14-3-3 (Fig. 7A). This region, which is highly conserved across RGS proteins, contains a 14-3-3 binding site in a subset of RGS proteins, including RGS3, RGS7, and RGS8 (Fig. 7, A and B). Alignment of several representative RGS proteins (ClustalW, www.ebi.ac.uk/ clustalw/) reveals a highly conserved SY/F/L motif followed by either a proline (or serine/threonine) or arginine/alanine/lysine. A mutation of Pro-to-Arg at position 436 changed RGS7 from a 14-3-3 binding protein to an RGS4-type protein that failed to bind 14-3-3 (Fig. 7B). The ability of RGS7 P436R to accelerate the deactivation of GIRK1/2 (Fig. 7C) was no longer inhibited by 14-3-3 (Fig. 7D). These findings demonstrate that the activity of RGS7 is regulated in vivo by binding of 14-3-3 to a conserved SYP motif. The mutational analysis of RGS4 and RGS7 is consistent with the 14-3-3 binding motifs defined by FIG. 6. The RGS4-mediated acceleration of GIRK deactivation is not affected by 14-3-3. A, Xenopus oocytes were microinjected with cRNA encoding for GIRK1 and GIRK 2, M2 AChR, and RGS4 (or water in control oocytes) as indicated. Depicted is the current decrease fit of a control oocyte and an oocyte expressing RGS4, revealing a substantial change in deactivation kinetics in the presence of RGS4. B, comparison of deactivation time constants of control oocytes and oocytes expressing RGS4. Expression of RGS4 significantly reduced the deactivation time constant for the deactivation of actetylcholine-evoked currents. C, microinjection of neither GST nor GST.14-3-3 altered the RGS4-mediated changes of GIRK1/2 deactivation kinetics. n.s., not significant.
RGS7 belongs to a subfamily of RGS proteins that contain a characteristic G␥-like domain, which forms obligate heterodimers with the G␤ subunit G␤ 5 (23)(24)(25). G␤ 5 co-expression dramatically enhances the activity of RGS7; in oocytes, coexpression accelerates the activation and deactivation of GIRK channels (11). In vitro, the RGS domain of RGS7 efficiently accelerates GTP hydrolysis of either G␣ i or G␣ o (26); in vivo, the RGS7⅐G␤ 5 complex appears to predominantly inhibit G␣ odependent signaling (27). RGS7 exists as both unmodified cytosolic and palmitoylated membrane-derived forms in the brain (28). Both forms of RGS7, when complexed with G␤ 5 , are equally effective stimulators of G␣ o GTPase activity, suggesting that palmitoylation of RGS7 occurs outside of the RGS domain and does not prevent RGS7/G␣ o interactions. RGS7 is rapidly degraded by the ubiquitin-dependent proteasome (29), a process that is inhibited by TNF-␣ (1). The TNF-mediated accumulation of RGS7 requires activation of the MAP kinase p38 and the presence of a short serine/threonine-rich motif (Ser 241 , Thr 245 , Thr 247 ) located upstream of the G␥-like domain of RGS7 (1), phosphorylation of which inhibits the degradation of RGS7. In contrast, phosphorylation of serine 434 mediates association of RGS7 with 14-3-3, which inhibits the GAP activity of RGS7 in vitro (6). Consistent with in vitro results, 14-3-3 significantly slowed the deactivation of GIRKs in oocytes in the presence of RGS7, suggesting that the interaction between 14-3-3 and RGS7 is an important regulatory mechanism of RGS7 activity in vivo. The interaction between RGS7 and 14-3-3 appears to be constitutive, i.e. in resting cells most of the cellular RGS7 is phosphorylated on serine 434 and complexed with 14-3-3 (6). To monitor the changes in phosphorylation state of serine 434, we generated a phospho-specific antiserum and analyzed serine 434 phosphorylation in response to extracellular stimulation in transfected cells. Our results demonstrate that cellular activation by TNF-␣ inhibits phosphorylation of serine 434 and thereby releases RGS7 from its inhibitory interaction with 14-3-3. Hence, TNF-␣ rapidly increases the concentration of functionally active RGS7 protein through two mechanisms. TNF-induced dephosphorylation of serine 434 liberates RGS7 from 14-3-3 binding and inhibition. In addition, TNF-␣ augments RGS7 levels through p38-dependent phosphorylation (1). The onset of both dephosphorylation and accumulation of RGS7 occur rapidly (15-30 min) after stimulation with TNF-␣. However, TNF-mediated dephosphorylation and dissociation with 14-3-3 peaks within 2 h, whereas accumulation of RGS7 peaks at 4 -6 h in transfected cells, the endogenous cell line, and mouse brain. The 14-3-3 binding motif typically contains a phosphorylated serine followed by a proline at position ϩ2 relative to serine. Although the proline at position ϩ2 is the preferred amino acid, certain other residues can promote binding of 14-3-3, including Gly, Gln, Asn, Phe, Ser, and Thr (30). All other residues (most notably Arg, Lys, and Ala) abrogate the interaction with 14-3-3. As predicted from high resolution x-ray structure and peptide library screens, a proline to arginine substitution at position 436 of RGS7 completely abrogated binding of 14-3-3 (31). Based on the amino acid at position ϩ2 relative to serine 434 of RGS7, RGS proteins can thus be divided into two categories, 14-3-3binding RGS proteins and RGS proteins with an arginine, alanine, or lysine at position ϩ2, which effectively abrogates 14-3-3 binding. Serine 434 of RGS7, located within one of the three sites mediating interaction with G␣ subunits, is highly conserved among all mammalian RGS proteins and is likely to be critical for the interaction with G␣ subunits and the function of RGS proteins. It is therefore not surprising that phosphorylation of this serine residue inhibits the GAP activity of RGS7 in vitro (6). The amino acid at position ϩ2 does not directly participate in G␣ binding, is less well conserved and appears to be the crucial determinant for 14-3-3 binding of RGS proteins. To support this hypothesis, we mutated the proline at position ϩ2 of RGS7 into arginine (RGS7 P436R ). Using the RGS7 P436R mutation in oocytes expressing GIRK1/2, we showed that RGS7, while retaining its GAP activity, was no longer regulated by 14-3-3. Our data suggest that most RGS proteins are regulated by phosphorylation of serine 434, which interferes with binding to G␣ subunits. In a subgroup of RGS proteins, phosphorylation of this serine residue transforms this motif to an active 14-3-3 binding site with pronounced inhibition of GTPase activity in vitro (6) and in vivo. This is the first evidence that this interaction, and hence the activity of RGS proteins, can be regulated by extracellular signals. It remains to be examined how accumulation of active RGS7, following activation of p38 MAP kinase and release from its inhibitory interaction with 14-3-3, alters G protein-mediated signaling in native tissues. Since RGS7 is highly expressed in the brain, we postulate that transient accumulation of functionally active RGS7 has an important role in modulating the activation and deactivation kinetics of neuronal GIRKs and thus in determining how neuronal cells react to repetitive signals and participate in complex neuronal networks.